OPINION
BUTZNER, Senior Circuit Judge:As a result of a collision in the Chesapeake Bay, the district court, sitting in admiralty, entered judgment on the complaint of the owners and insurance carrier of the tug Patricia against the Spica Corporation and the. tug Spica for 75% of the Patricia’s damages. The court entered judgment on Spica’s counterclaim in the amount of 25% of its damages. We affirm.
I
Because the district court’s factual findings and its determinations of credibility are essential to understanding its judgment, we set forth this part of its opinion verbatim:
Initially, the Court notes that the parties present markedly different versions of some portions of the same incident. Therefore, in the Court’s recitation of the facts, it will indicate where the parties’ positions diverge.
On October 12, 1989, which was a clear night, the Spica and its crew were hauling dredge spoils from the dredge “Virginian” down to a dump spot near Poole’s Island. On the same evening, the Patricia was returning with its crew in a westerly direction down the Chesapeake Bay. At the helm of the Patricia at the time of the collision was Stephen Waters, who had some experience navigating in the waters where the collision occurred. At the helm on the Spica was Mark Young, who had limited experience with navigation and piloting a tug.
At about 1 a.m., as the Patricia was moving in a westerly course down the Chesapeake, she left the deep water shipping channel just off Turkey Point at buoys 7 and 8 and began to cut directly across the flats in a straight line toward *195buoys 43 and 44 off Howell Point.1 Meanwhile, the Spiea was coming up on the red side, or southern side, of the channel just north of Poole’s Island, and then cut across the channel to continue its journey parallel to the channel and outside of it on the green side, or northern side. Each pilot spotted the other far in the distance approximately 30 minutes prior to the collision.
A
According to Waters, as the Patricia proceeded on her course toward a midpoint between buoys 43 and 44, the Spica, which was still a considerable distance away, appeared about 10 degrees off the Patricia’s starboard bow. This relative position of the two tugs remained constant until the Patricia reached the channel just off Howell Point. Waters confirmed using his radar equipment that the Spica was outside the green side of the channel, just west of buoys 41 and 42. Waters saw both the red and green lights of the Spica. He anticipated that the two tugs would pass each other on their starboard sides.
When the Patricia was abeam buoys 43 and 44, Waters turned the Patricia slightly to the right in order to stay on the green side of the channel. Waters stated that the turn was a mild right turn of about two degrees. At that point, the Spica was approximately two miles away and about five degrees off the Patricia’s starboard bow. Waters believed that the two tugs would pass within one quarter mile of each other, which he indicated was closer than he liked. He noted that the Spica seemed to be edging toward the channel on the green side. Consequently, with the Spica about one and one half miles away, Waters turned the Patricia approximately five degrees to port in order to give the Spica more room on a starboard-to-starboard pass. Waters could see only the green light of the Spica after his turn to port. The Spiea continued to be about five degrees off the Patricia’s starboard bow.
Suddenly, when the two tugs were between an eighth and a quarter of a mile apart, Waters noticed that the Spica’s lights were changing from green to red, which indicated that the Spica was turning into the channel in front of the Patricia. Waters attempted to avoid the collision, but was unable to do so. The Spica hit him on the starboard barge at a ninety degree angle.
B
According to Young, at the time of the collision, the Spiea was returning to the dredge Virginian to load up more dredge spoils for dumping. He stated that the standard course was for the Spica to proceed up the bay outside of the green side of the channel and then either to cut back across the channel toward the dredge, or to continue past the dredge outside the channel and cut back. When the Spica was within one mile of the dredge, Young intended to awaken the captain, James Widgeon, before docking with the dredge. The mud scow was lashed to the port side of the Spica and caused a blind spot in Young’s forward view.
Young steered the Spica outside of the green side of the channel within 50 to 100 feet of the channel markers. He noted that both the Spiea and Patricia were running outside the green. He did not alert Widgeon to any problem and stated that he believed the tugs were set for a port-to-port pass.
Finally, to assess the Patricia’s position, Young used his spotlight. At this time he states that the Patricia was turning in front of him, showing the port side of its barge.2 Thus, his testimony was that the Patricia was still outside the green side of the channel and turning to port in front of him across his bow. Young then immediately reversed his engines but was unable *196to avoid the collision. The Spiea’s mud scow hit the Patricia’s starboard barge at a ninety degree angle.
Widgeon testified that Young stated to him two or three days after the accident that he had thought the Patricia was off to his port and moving across the flats and that Young had lost sight of the Patricia. He told Widgeon that the next thing he became aware of was that the Patricia was turning across his bow.
C
The two versions of the facts presented by the parties are sharply at odds in regard to the relative positions of the tugs just before the collision, and the Court cannot reconcile the two versions. According to Waters, the Patricia was operating in the channel and to the starboard of the Spica; whereas, according to Young, the Patricia was outside the channel and to the port of the Spica. Based on the trial testimony of Waters and Widgeon, as well as the deposition testimony of Young and Waters submitted by the parties, the Court finds that Waters’ testimony is credible and represents accurately the events leading up to the collision.
C.G. Willis, Inc. v. Spica Corp., No. B-90-889, slip op. at 2-6 (D.Md. June 17, 1992).
II
Admitting its own negligence, Spica nevertheless contends that the district court erred by not attributing the majority of fault to the Patricia. It insists that Patricia’s violation of the Inland Navigational Rules and its failure to avoid the risk of collision were the principal causes of the collision. Spica’s primary contention is that the Patricia did not maneuver according to rule 14, 33 U.S.C. § 2014, which governs situations where there is a risk of a head-on collision and which states that when such a situation exists a vessel “shall alter her course to starboard so that each shall pass on the port side of the other.” Spica contends that Waters’ decision to turn the Patricia slightly to port when the two tugs were one and one half miles apart violated this rule. Spica also contends that the failure of the Patricia to post a lookout, to use its radar, and to signal by whistle that it contemplated a starboard-to-starboard passage contributed to the cause of the collision.
The Patricia contends that no head-on meeting occurred, and, therefore, it was not required to turn to starboard. It also asserts that the absence of a lookout and failure to use radar could not be the cause of the collision. The cause of the collision, it insists, was the act of the Spica in blindly turning to starboard after it lost sight of the Patricia due to the obstruction on its barge.
We review for clear error under Fed.R.Civ.P. 52(a) the district court’s finding of negligence, its partial exoneration of the Patricia, and its apportionment of liability. McAllister v. United States, 348 U.S. 19, 20-21, 75 S.Ct. 6, 8, 99 L.Ed. 20 (1954); United States Fire Ins. Co. v. Allied Towing Corp., 966 F.2d 820, 823 (4th Cir.1992). We review de novo the district court’s interpretation of statutes. McDermott International, Inc. v. Wilander, 498 U.S. 337, 356, 111 S.Ct. 807, 818, 112 L.Ed.2d 866 (1991).
Ill
Both the Spica and the Patricia were under a duty to avoid a risk of collision. If there is any doubt, a risk of collision shall be deemed to exist. Furthermore, a risk should be deemed to exist if the compass bearing of an approaching vessel does not appreciably change. A risk also may exist when approaching a vessel at close range. Rule 7, 33 U.S.C. § 2007. “[I]t is the risk of collision, not the collision itself, that masters must avoid.” Ocean S.S. Co. v. United States, 38 F.2d 782, 784 (2d Cir.1930).
Rule 14 requires vessels meeting head-on to alter their course to starboard for port-to-port passage. A head-on situation shall be deemed to exist when at night a vessel sees both side lights of an approaching vessel.
On Spica’s charge that the Patricia violated rule 14, the district court’s findings are specific:
The Court notes that neither Waters nor Young ever believed that a head-on collision was imminent. On the contrary, Wa*197ters testified that until just before the collision, he always believed that the tugs would make a starboard-to-starboard pass. While it is true that Waters moved to port in the channel because the tugs were closer than he liked, it does not follow that a head-on situation had developed.
Nor does the fact that Waters at one time saw both the red and green lights of the Spica indicate a head-on meeting. Waters’ testimony demonstrates that when the two tugs were still some distance apart, he did see both lights on the Spica. However, after moving to port in the channel to give the Spica more room to pass, Waters stated that he could see only the green lights on the Spica.
C.G. Willis, No. B-90-889, slip op. at 7 (emphasis in original). These findings are amply supported by the evidence. Because they are not clearly erroneous, Rule 52(a) counsels their acceptance. The risk of collision that could have occurred after the Patricia turned to port was solely from an unexpected change of course by one of the tugs. And this is actually what happened. The Spica, having lost sight of the Patricia, turned starboard, not to pass the Patricia port-to-port, but to make her way to the dredge she was serving. The district court did not err in finding that the Spica’s starboard turn, not the Patricia’s slight turn to port, was the primary cause of the collision.
In agreement with the district court, we conclude that the Patricia was not at fault as a matter of law. The district court rightly rejected the Spica’s argument that a head-on situation existed as a matter of law because the Spica was only 10 degrees, or less, off the Patricia’s starboard bow.
In support of its argument, Spica cites In re Ocean Foods Boat Co., 692 F.Supp. 1253, 1261 (D.Or.1988), which involved a collision in the open ocean arising out of a crossing situation. The court observed that as a general rule differences between two vessels of less than 11% degrees are treated as a head-on situation. Ocean Foods does not purport to apply this general rule to the marked channels of inland waters such as the Chesapeake Bay. John V. Noel, Jr., Knight’s Modern Seamanship 579 (17th ed.1984), explains the difference between a starboard-to-starboard meeting in inland waters and a head-on situation as follows:
24.16 Starboard-to-Starboard Meeting in Inland Waters (Not Within the Definition of Head-On) When vessels are meeting within the definition of “head-on,” a port-to-port passage is required. The Inland Rules prescribe signals for a starboard-to-starboard passing and require that the signals be exchanged whenever the CPA [Closest Point of Approach] is within one-half mile. In this situation, each vessel will have only the green sidelight of the other in sight.
The Patricia’s helmsman intended a starboard-to-starboard passage. The Spica was off the Patricia’s starboard bow for a distance of one and a half miles, and the Patricia sighted only the Spica’s green light. One can draw a fair inference that if the Spica had a lookout, he would have seen the Patricia’s green light for a distance of one and a half miles. The situation resembles the starboard-to-starboard passage described in Knight’s Modem Seamanship, which refutes the notion that a head-on situation existed as a matter of law.
Patricia’s failure to signal a starboard-to-starboard passage violated rule 34(a), 33 U.S.C. § 2034(a), and for this omission the court appropriately sanctioned it. The court found that both vessels failed to signal their respective maneuvers by whistle in violation of Rule 34. It properly held that their statutory violations contributed to the collision, and it assessed the Patricia’s fault at 25%.
IV
Spica argues that the district court erred in assigning no fault to the Patricia for failing to have a special lookout posted in addition to the helmsman, while at the same time finding fault with the Spica for not having a proper lookout. In support of its position Spica cites Esso Standard Oil Co. v. Oil Screw Tug Maluco I, 332 F.2d 211 (4th Cir.1964), as well as Rule 5 of the Inland Navigational Rules, which requires that vessels post a proper lookout “so as to make a *198full appraisal of the situation and of the risk of collision.” 33 U.S.C. § 2005.
In Esso, a tanker and a tug collided on the Potomac River. The channel at the point of collision was only 200 feet wide, there was a jumble of red and green lights placed on the shore as well as on cofferdams extending into the river, some of which were installed but hours before the accident, and both vessels were navigating using blinding searchlights. Esso, 332 F.2d at 214. The two vessels collided. We held that in such a situation the operator of the tug was unable to fulfill the “requirements of a ‘proper lookout.’ ” Esso, 332 F.2d at 214.
Spica claims that this case and others establish the need for a special lookout as a matter of law. This is a misstatement of the requirement for a special lookout. In Esso, we stated that “if a special lookout was not needed, the tug must plead the adequacy of her general lookout in order to meet her burden.” Esso, 332 F.2d at 214. Commenting on Esso in China Union Lines, Ltd. v. A.O. Andersen & Co., 364 F.2d 769, 782 (5th Cir.1966), the court said: “If, in [Esso ], the judge had found that the pilot of the vessel saw everything which a bow pilot lookout could have seen, undoubtedly, the Fourth Circuit would not have reversed.”
Here, the Patricia met the burden of establishing that an additional lookout was unnecessary. Unlike in Esso, there were no confusing lights or blinding searchlights prohibiting the Patricia from following the progress of the Spica. The district court held that
the evidence is clear that [Patricia’s helmsman] never lost sight of the Spica. He knew where it was located from the moment he first spotted it and, in fact, altered course once to give it more sailing room. There simply is no basis on which the Court could find that an additional lookout on the Patricia would have prevented the collision.
C.G. Willis, Inc., No. B-90-889, slip op. at 12.
At the same time, the court found that the Spica’s helmsman had lost sight of the Patricia due to the forward obstruction caused by the mud scow and that his lack of knowledge regarding the Patricia’s whereabouts “caused him to turn to starboard into its side.” C.G. Willis, Inc., No. B-90-889, slip op. at 10. Here, as in Esso, the special lookout was needed on the Spica, a “myopic tug with a protruding and displaced tow.” Esso, 332 F.2d at 214. It was not required on the Patricia. See China Union Lines, 364 F.2d at 782.
V
Spica assigns error to the district court’s refusal to find any fault with the Patricia for not using radar plotting to track the Spica’s position, while faulting the Spica with not using its radar. The radar requirement is found in Rule 7, which mandates that vessels “use all available means appropriate to the prevailing circumstances and conditions to determine if risk of collision exists.” 33 U.S.C. § 2007(a). Radar should be used “to obtain early warning of risk of collision.” 33 U.S.C. § 2007(b).
The district court held:
[The Patricia’s helmsman] knew the position of the Spica and had it in sight. Even had he used radar plotting, he could not have anticipated that the Spica would take a sudden turn into the channel. Thus, the Court finds that [the Patricia’s] failure to use radar plotting could not have caused the accident.
C.G. Willis, Inc., No. B-90-889, slip op. at 13. At the same time, the court found that the collision was caused by the Spica’s failure to know where the Patricia was located. Radar plotting could have helped its helmsman know this position. The district court properly found the Spica at fault for not using its radar equipment.
Spica’s contention that radar plotting could have helped the Patricia detect the risk of collision is without merit. As the district court found, its helmsman never lost sight of the Spica. Radar plotting would have provided him with no additional information that could have allowed him to anticipate the Spica’s sudden turn into the Patricia’s side.
*199VI
The district court’s finding that the Spica was 75% at fault and the Patricia 25% is not clearly erroneous, and it is by this standard that we review this aspect of the court’s judgment. See United States Fire Ins. Co., 966 F.2d at 823. The court properly allocated the burden of proof by requiring each vessel to show that its statutory violations could not have at least contributed to the collision. See The Pennsylvania, 86 U.S. (19 Wall.) 125, 136, 22 L.Ed. 148 (1873).
AFFIRMED
. This opinion refers to the buoys by their current numbers. At the time of the accident buoys 7 and 8 were numbered 5 and 6, buoys 43 and 44 were numbered 37 and 38, and buoys 41 and 42 were buoys 35 and 36.
. This testimony is logically inconsistent with Young’s description of what occurred.